can any flatbeds be used upside down? Scanning oversized artwork. cis scanners

[Guest]

Anyone know of a mid price-range cis scanner that can be turned upside
down to scan oversized work? The scanners glass must actually be in
direct contact with work that is larger(wider and longer) than the
scanner glass.

the typical scanner has plastic framing higher than the glass surface
so when the scanner is placed on top of oversized work (like on a
table) the glass surface is not actually in contact with the work.

The hp4600 series of "see through" scanners would be a type of design
that I am looking for however they are low quality and have a ccd
sensor -not accurate enough for my projects. See the unique and
wonderful design here - I just wish they would make a cis model...
http://h10010.www1.hp.com/wwpc/us/en/sm/WF05a/15179-64195-215155-15202-215155-303776.html

Why must the work be in contact with the glass.. Because I need a cis
scanner and unlike ccd scanners the cis scanners have no depth of
field, cis scanners focus only at the glass surface. (If anyone knows
of a cis scanner that has depth of field >.5mm please let me know!)

Why do I want a cis scanner anyway? because dynamic range and overall
color quality is of low importance for my projects however geometric
accuracy is of the highest importance. Cis scanners are very
geometrically accurate by virtue of their architecture and appropriate
for my project. The mid range ccd scanners are not geometrically
accurate enough for my projects(I have tried many) and I don't want to
spend several thousands of dollars to buy a large platform high end ccd
scanner. I also need to be mobile with the scanner so it needs to be
somewhat durable and small.

this is one article that demonstrates the accuracy of even low end cis
scanners and the typical problems with ccd scanners.
http://195.11.224.13/pdf/wpcisccd.pdf

my projects ... scanning very detailed oversized (28"x48"&+) work that
is on rigid backings(like glass and wood) and then stitching the images
together with a quality stitching program or manually in Photoshop.

 

[Wayne]

The hp4600 series of "see through" scanners would be a type of design
that I am looking for however they are low quality and have a ccd
sensor -not accurate enough for my projects.

Curious why you assume it is CCD? My assumption is the opposite, because it
is only 0.7 inches thick, insufficient space to contain the lens and mirrors
for CCD. The document you referenced says "CCD scanner cannot be made this
thin". I think there is none thinner, so to me, it seems to be CIS by
definition.

(If anyone knows
of a cis scanner that has depth of field >.5mm please let me know!)

Dont hold your breath waiting... that's one advantage of the CCD lens. CCD
depth of field is not greatly more than zero, but not equal to zero like the
CIS. The CIS sensor is simply placed up close to the glass. It probably
could be done differently, but it isnt, the point is avoid the expensive of a
lens focusing system.

this is one article that demonstrates the accuracy of even low end cis
scanners and the typical problems with ccd scanners.
http://195.11.224.13/pdf/wpcisccd.pdf

Your referenced document seemed to ignore concern for the vertical dimension
size accuracy. The lens magnification of CCD units can be slightly off, but
the vertical dimension (of both types) is dependent on the motor and gears and
flexible drive belts.

The 6.027 inches seems a minor problem, because all you actually have is
pixels, and it is easy to scale the image to print any size you wish, half
size, double size, or 6.00 inches size (but it's not inconceivable the printer
motors could have similar concerns)

 

[Terry]

Curious why you assume it is CCD? My assumption is the opposite, because it
is only 0.7 inches thick, insufficient space to contain the lens and mirrors
for CCD. The document you referenced says "CCD scanner cannot be made this
thin". I think there is none thinner, so to me, it seems to be CIS by
definition.


Dont hold your breath waiting... that's one advantage of the CCD lens. CCD
depth of field is not greatly more than zero, but not equal to zero like the
CIS. The CIS sensor is simply placed up close to the glass. It probably
could be done differently, but it isnt, the point is avoid the expensive of a
lens focusing system.


Your referenced document seemed to ignore concern for the vertical dimension
size accuracy. The lens magnification of CCD units can be slightly off, but
the vertical dimension (of both types) is dependent on the motor and gears and
flexible drive belts.

The 6.027 inches seems a minor problem, because all you actually have is
pixels, and it is easy to scale the image to print any size you wish, half
size, double size, or 6.00 inches size (but it's not inconceivable the printer
motors could have similar concerns)

Try this link:
http://consumer.usa.canon.com/ir/controller?act=ModelTechSpecsAct&fcategoryid=120&modelid=11011
Its a CIS scanner from Canon with a double hinge so you can scan in
almost any position you want.

 

[CSM1]

Anyone know of a mid price-range cis scanner that can be turned upside
down to scan oversized work? The scanners glass must actually be in
direct contact with work that is larger(wider and longer) than the
scanner glass.

the typical scanner has plastic framing higher than the glass surface
so when the scanner is placed on top of oversized work (like on a
table) the glass surface is not actually in contact with the work.

The hp4600 series of "see through" scanners would be a type of design
that I am looking for however they are low quality and have a ccd
sensor -not accurate enough for my projects. See the unique and
wonderful design here - I just wish they would make a cis model...
http://h10010.www1.hp.com/wwpc/us/en/sm/WF05a/15179-64195-215155-15202-215155-303776.html

Why must the work be in contact with the glass.. Because I need a cis
scanner and unlike ccd scanners the cis scanners have no depth of
field, cis scanners focus only at the glass surface. (If anyone knows
of a cis scanner that has depth of field >.5mm please let me know!)

Why do I want a cis scanner anyway? because dynamic range and overall
color quality is of low importance for my projects however geometric
accuracy is of the highest importance. Cis scanners are very
geometrically accurate by virtue of their architecture and appropriate
for my project. The mid range ccd scanners are not geometrically
accurate enough for my projects(I have tried many) and I don't want to
spend several thousands of dollars to buy a large platform high end ccd
scanner. I also need to be mobile with the scanner so it needs to be
somewhat durable and small.

this is one article that demonstrates the accuracy of even low end cis
scanners and the typical problems with ccd scanners.
http://195.11.224.13/pdf/wpcisccd.pdf

A well designed sales pitch for a CIS scanner, Probably Canon.
I did not find the test for scanning a ruler a accurate demonstration that a
CCD is any less accurate in the linear direction, (did you turn the ruler 90
degrees?)

For geometrically accuracy, you should include a scale (a ruler or caliper)
in the edge of the scan for comparison. (Put the scale in the overlap area).

No scanned image has an inherent size. Only a few file formats include the
DPI an image is scanned at, which you must know to reproduce the same size
print. With a scale in the image, one can resize the image in any photo
editor to make the scale the same size. If you are trying for 1/1000 inch
accuracy, you do not want to use a consumer grade scanner.

See somebody that has developed the technology and has done the project.
http://graphics.stanford.edu/projects/mich/

Also see:
http://www.ifla.org/IV/ifla61/61-murj.htm
The image links seem to be broken (images not there?).

For further research, Google for "scanning project art museum" without the
quotes.

my projects ... scanning very detailed oversized (28"x48"&+) work that
is on rigid backings(like glass and wood) and then stitching the images
together with a quality stitching program or manually in Photoshop.

You would probably be better off photographing the art in a photographic
studio with proper lighting and a medium format (120 film) or large format
film camera such as 4" x 5" or 8" x10". Then scan the Film. Scanners for 4 x
5 film are available at around $400 and up.
http://consumer.usa.canon.com/ir/controller?act=ModelDetailAct&fcategoryid=120&modelid=10446

Epson makes a scanner that will scan up to 8" X 10" transparencies. About
$600.
http://www.epson.com/cgi-bin/Store/consumer/consDetail.jsp?BV_UseBVCookie=yes&oid=49164281

I have photographed printed circuit board layouts on 35 mm high contrast
film and produced masks for etching Printed Circuit boards with at least
1/100 inch accuracy. All I did was include a scale in the image. When
developing the print I used the scale in the darkroom and scaled the image
to match the ruler.

 

[Guest]

Curious why you assume it is CCD? My assumption is the opposite, because it

is only 0.7 inches thick, insufficient space to contain the lens and mirrors
for CCD. The document you referenced says "CCD scanner cannot be made this
thin". I think there is none thinner, so to me, it seems to be CIS by
definition.

exactly what I thought - however
1) in reviews of the 4600 people were talking about long lamp warm up
times before the scanner would allow a scan - this could be just a
software issue.
2) HP's data sheet for the 4600 says it is a ccd - see
http://www.shopping.hp.com/webapp/s...rs&subcat1=mid_range&catLevel=3#defaultAnchor
click on view all specks or download pdf.
3) still I thought it might be a mistake in the documentation and the
software might unnecessarily be 'warming up' the bulb/lights. If the
software was adapted from a ccd scanner HP may have unnecessarily
engineered a warm up time on a cis scanner when they didn't need
to... SO I called HP tec support team and questioned them in depth
about this product. They worked hard to find out the answer, 15 min
back and forth with me on the phone, several people involved as a team
researching the question... HP says that it is in fact a ccd scanner.
I still have my doubts for the reasons Wayne stated and I may just go
buy one to find out for sure.

The 6.027 inches seems a minor problem, because all you actually have is
pixels, and it is easy to scale the image to print any size you wish, half
size, double size, or 6.00 inches size (but it's not inconceivable the printer
motors could have similar concerns)

TURE! Initially I did try to scale the images however the real problem
is that the horizontal dimension is not just stretched or shrunk it is
also distorted i.e. the stretching or shrinking is NOT linear over the
horizontal width, there are spots of accelerated shrinking or
stretching or distortion. Horizontally scaling generally helped and
was more accurate but it didn't fix the distortion. To accommodate the
distortion I scaled an image in several sections across the horizontal
dimension and recorded the different scaling for each horizontal
section. I figured this would give me a 'formula' I could use to
adjust all scans made from the specific scanner to correct for the
distortions specific to that scanner. To test this I made several more
exact scans and scaled the horizontal sections with the recorded
scalings or 'formula' from the first scan... to my dismay the
scaled scans were way off after this scaling and this method did not
create accurately scaled scans. My conclusion is that the distorted
spots morph, change, move or jump around from scan to scan - I think
the heat caused by the lamp and the cooling from movement during
scanning is causing the mirrors and plastic housings to expand and
contract causing a slight distortion from scan to scan. Also the
movement of the motor and carriage during scans may jiggle the mirrors
and ccd a bit. I didn't want to include all of that in the original
post because of length. Seems like cis scanners with cooler led lights
that only are on during the scan and the cis sensor configuration
across the whole horizontal dimension without mirrors etc gets around
this whole issue.

Your referenced document seemed to ignore concern for the vertical dimension
size accuracy. The lens magnification of CCD units can be slightly off, but
the vertical dimension (of both types) is dependent on the motor and gears and
flexible drive belts.

I didn't mention it in my original post but In my tests of several
scanners the vertical dimension was extremely accurate across the whole
scan, without any measurable distortion, in all of the scanners I
tested. It seems that the stepper motors, drive systems and vertical
dimension has been perfected, even on the low end scanners. To bad
they don't make a mid range scanner with a one dot sensor moved both
vertically and horizontally with stepper motors.. scans would take
forever but seems like they would be extremely accurate.

 

[Wayne]

3) still I thought it might be a mistake in the documentation and the
software might unnecessarily be 'warming up' the bulb/lights.

Interesting. You have definitely convinced me that the specs do say CCD.
I cant say I am a believer, but I have not seen one.

the real problem
is that the horizontal dimension is not just stretched or shrunk it is
also distorted i.e. the stretching or shrinking is NOT linear over the
horizontal width, there are spots of accelerated shrinking or
stretching or distortion.

Also strange. Then scanning a sheet of quadrille paper must give an
unacceptable result. I assume your stitching a dozen images together is not
the issue. If it has a lens, it can only be a really tiny lens - maybe it
could be barrel or pincushion distortion. Or if a full length sensor (CIS
style instead of a lens), its array might be crudely constructed (to do what
you described). Either way would seem a moot point in this case however,
doesnt sound like it can do the job.

movement of the motor and carriage during scans may jiggle the mirrors
and ccd a bit.

If actually present, lamp, mirrors, lens and CCD would be all be attached to
the carriage, which is what the carriage is. Everything just rides along as
a unit, required so that the focused distance was constant. Normally any
little issue is that the carriage guide rods may not be perfectly parallel to
the bed, giving a slight skew over the full path.

I'm sorry, I have no clue about it, but it would seem your needs are more
critical than a $50 bottom-end scanner.

You said your goal was scanning 28x48 inch work, but that image quality wasnt
much issue, only the dimensional geometry. Unless you also have large size
printing requirements, you might consider a digital camera with tripod or
fixed mount, and a well lighted copy setup. The image will be greatly
smaller, but 6 megapixels full frame should print 8x12 inches, and will be
better if paintings, and less good for line drawings. Or it could be a film
camera, and you could scan the film. Large sheet film if you need to print a
similar size.

 

[CSM1]

Interesting. You have definitely convinced me that the specs do say CCD.
I cant say I am a believer, but I have not seen one.


Also strange. Then scanning a sheet of quadrille paper must give an
unacceptable result. I assume your stitching a dozen images together is
not
the issue. If it has a lens, it can only be a really tiny lens - maybe it
could be barrel or pincushion distortion. Or if a full length sensor (CIS
style instead of a lens), its array might be crudely constructed (to do
what
you described). Either way would seem a moot point in this case however,
doesnt sound like it can do the job.


If actually present, lamp, mirrors, lens and CCD would be all be attached
to
the carriage, which is what the carriage is. Everything just rides along
as
a unit, required so that the focused distance was constant. Normally any
little issue is that the carriage guide rods may not be perfectly parallel
to
the bed, giving a slight skew over the full path.

I'm sorry, I have no clue about it, but it would seem your needs are more
critical than a $50 bottom-end scanner.

You said your goal was scanning 28x48 inch work, but that image quality
wasnt
much issue, only the dimensional geometry. Unless you also have large
size
printing requirements, you might consider a digital camera with tripod or
fixed mount, and a well lighted copy setup. The image will be greatly
smaller, but 6 megapixels full frame should print 8x12 inches, and will be
better if paintings, and less good for line drawings. Or it could be a
film
camera, and you could scan the film. Large sheet film if you need to
print a
similar size.

Maybe this is what you need.
http://www.paradigmimaging.com/services/fine_art_scanning.htm

The page with the Large Format scanners.
http://www.scantopia.com/Scanner/scanners.html

 

[Warren Weber]

Anyone know of a mid price-range cis scanner that can be turned upside
down to scan oversized work? The scanners glass must actually be in
direct contact with work that is larger(wider and longer) than the
scanner glass.

the typical scanner has plastic framing higher than the glass surface
so when the scanner is placed on top of oversized work (like on a
table) the glass surface is not actually in contact with the work.

The hp4600 series of "see through" scanners would be a type of design
that I am looking for however they are low quality and have a ccd
sensor -not accurate enough for my projects. See the unique and
wonderful design here - I just wish they would make a cis model...
http://h10010.www1.hp.com/wwpc/us/en/sm/WF05a/15179-64195-215155-15202-215155-303776.html

Why must the work be in contact with the glass.. Because I need a cis
scanner and unlike ccd scanners the cis scanners have no depth of
field, cis scanners focus only at the glass surface. (If anyone knows
of a cis scanner that has depth of field >.5mm please let me know!)

Why do I want a cis scanner anyway? because dynamic range and overall
color quality is of low importance for my projects however geometric
accuracy is of the highest importance. Cis scanners are very
geometrically accurate by virtue of their architecture and appropriate
for my project. The mid range ccd scanners are not geometrically
accurate enough for my projects(I have tried many) and I don't want to
spend several thousands of dollars to buy a large platform high end ccd
scanner. I also need to be mobile with the scanner so it needs to be
somewhat durable and small.

this is one article that demonstrates the accuracy of even low end cis
scanners and the typical problems with ccd scanners.
http://195.11.224.13/pdf/wpcisccd.pdf

my projects ... scanning very detailed oversized (28"x48"&+) work that
is on rigid backings(like glass and wood) and then stitching the images
together with a quality stitching program or manually in Photoshop.

If I want to copy something larger than my scanner will cover I just
photograph it with a high megabyte camera and put on computer and go from
there. Warren